STRESS INDUCED FACTOR 2, a Leucine-Rich Repeat Kinase Regulates Basal Plant Pathogen Defense^,

Arabidopsis SIF2 belongs to the stress-responsive SIF protein kinase family and regulates basal defense to pathogen infection by modulating the PTI-mediated signal transduction pathway.

Abstract

Protein kinases play fundamental roles in plant development and environmental stress responses. Here, we identified the STRESS INDUCED FACTOR (SIF) gene family, which encodes four leucine-rich repeat receptor-like protein kinases in Arabidopsis (Arabidopsis thaliana). The four genes, SIF1 to SIF4, are clustered in the genome and highly conserved, but they have temporally and spatially distinct expression patterns. We employed Arabidopsis SIF knockout mutants and overexpression transgenics to examine SIF involvement during plant pathogen defense. SIF genes are rapidly induced by biotic or abiotic stresses, and SIF proteins localize to the plasma membrane. Simultaneous knockout of SIF1 and SIF2 led to improved plant salt tolerance, whereas SIF2 overexpression enhanced PAMP-triggered immunity and prompted basal plant defenses, significantly improving pathogen resistance. Furthermore, SIF2 overexpression plants exhibited up-regulated expression of the defense-related genes WRKY53 and flg22-INDUCED RECEPTOR-LIKE KINASE1 as well as enhanced MPK3/MPK6 phosphorylation upon pathogen and elicitor treatments. The expression of the calcium signaling-related gene PHOSPHATE-INDUCED1 also was enhanced in the SIF2-overexpressing lines upon pathogen inoculation but repressed in the sif2 mutants. Bimolecular fluorescence complementation demonstrates that the BRI1-ASSOCIATED RECEPTOR KINASE1 protein is a coreceptor of the SIF2 kinase in the signal transduction pathway during pathogen invasion. These findings characterize a stress-responsive protein kinase family and illustrate how SIF2 modulates signal transduction for effective plant pathogenic defense.

Environmental stresses, such as drought, salinity, and pathogens, are among the most important factors influencing agricultural production. An understanding of the molecular mechanisms underlying plant responses to adverse environmental conditions is critical for developing effective genetic improvement strategies for crop species.

In response to abiotic stresses, plants evolved multiple defense mechanisms and strategies. For example, upon drought stress, plants are capable of accelerating their life cycle and fruiting before severe water deficit symptoms occur. Plants also can adopt avoidance and tolerance strategies during drought or salinity stress: stomata are closed to prevent water loss, osmolytes such as Pro are synthesized to maintain appropriate osmotic pressure, the expression of transporter genes is regulated to help plants exclude or compartmentalize harmful ions, and root growth is greatly promoted to maximize water uptake (Chaves et al., 2003; Shkolnik-Inbar et al., 2013).

In response to biotic stress caused by pathogens, plants utilize two layers of innate immunity (Jones and Dangl, 2006). The PAMP-triggered immunity (PTI) pathway is the first layer of the plant immunity system and involves the recognition of pathogen-associated molecular patterns (PAMPs), such as flagellin or elongation factor Tu. However, PTI can be repressed by effectors injected into plant cells by some pathogen species through a type III secretion system. To counter pathogens at this level, effector-triggered immunity, the second layer of plant immunity, is initiated to suppress these effectors and resist the pathogen infection (Jones and Dangl, 2006).

The plasma membrane offers plant cells a stable and orderly protoplasm environment that is isolated from the external environment (Serrano, 1984; Laude and Prior, 2004). To counter stress and survive adverse environments, cells need to receive and transduce extracellular stress signals into the intracellular environment through the plasma membrane barrier. Many membrane-anchored proteins, such as receptor-like protein kinases (RLKs), act as sensors and receptors mediating signaling transduction. Leucine-rich repeat (LRR)-RLKs are composed of the largest subfamily of transmembrane receptor-like protein kinases in Arabidopsis (Arabidopsis thaliana; Torii, 2004; Gou et al., 2010; Hok et al., 2011). Over the past 20 years, plant LRR-RLKs were found to play fundamental roles in cell proliferation, photomorphogenesis, and biotic and abiotic stress responses (Deeken and Kaldenhoff, 1997; Li and Chory, 1997; Fletcher et al., 1999; Xiang et al., 2006; de Lorenzo et al., 2009; Antolín-Llovera et al., 2012). A Medicago truncatula LRR-RLK gene, SRLK, was proven to be a possible receptor, which functions in plant resistance against salt stress (de Lorenzo et al., 2009). RECEPTOR-LIKE PROTEIN KINASE1 (RPK1), an Arabidopsis LRR-RLK, is strongly up-regulated under abiotic stress and abscisic acid (ABA) treatment (Hong et al., 1997). An Arabidopsis line overexpressing RPK1 exhibited enhanced salt tolerance, indicating the important function of RPK1 in abiotic stress resistance (Osakabe et al., 2010). So far, only a few LRR-RLKs, such as FLAGELLIN SENSITIVE2 (FLS2), EF-Tu RECEPTOR (EFR), PEP RECEPTOR1/2 (PEPR1/2), and BRI1-ASSOCIATED RECEPTOR KINASE1 (BAK1), have been identified to function in signal transduction upon pathogen invasion (Chinchilla et al., 2007; Postel et al., 2010; Schulze et al., 2010). Of particular importance, BAK1 is a multifunction LRR-RLK in Arabidopsis. Besides its critical role in the perception of brassinosteroid, BAK1 also is an important coreceptor in the Arabidopsis PTI pathway. A previous study showed that, after FLS2 binds flagellin, the C-terminal domains of the BAK1 LRR immediately forms a sandwich structure with the C terminus of flagellin and the FLS2 LRR domains (Li et al., 2002; Sun et al., 2013). The conformational changes caused by this BAK1-flagellin-FLS2 sandwich structure promote BAK1 to autophosphorylate its own kinase domain and transphosphorylate the kinase domain of FLS2, activating FLS2 and/or BAK1 to recruit and trigger downstream signaling cascades (Schwessinger et al., 2011). Mutations in critical amino acid residues of the BAK1 LRR domains attenuate both interactions and phosphorylation between this FLS2 and BAK1 heterodimer, and a mutation in the BAK1 kinase domain negatively impacts its phosphorylation ability (Schwessinger et al., 2011; Sun et al., 2013). Many LRR-RLKs act as receptors in the PTI pathway, recognizing external PAMP elicitors and heterodimerizing with BAK1 to trigger internal signaling transduction.

We have identified an LRR-RLK family, annotated as SIF (for stress-induced factor), using bioinformatics analysis with Arabidopsis cDNA microarray data. Here, we demonstrate that the four SIF family members may participate in different stress response signaling transduction pathways in Arabidopsis, although their highly conserved sequences suggest that they may have similar functions. Using a sif2 T-DNA insertion mutant and SIF2-overexpressing lines, we determined that SIF2 is a critical element in the PTI pathway and positively regulates plant basal defenses against pathogens. The evidence from our research indicates that SIF2 interacts with BAK1 upon pathogen infection to recruit and activate the downstream MAPK cascade, inducing the expression of WRKY53 and flg22-INDUCED RECEPTOR-LIKE KINASE1 (FRK1) and triggering basal defense responses. The quantitative reverse transcription PCR (RT-qPCR) results also show that, upon pathogen infection, the induced expression of the calcium signaling gene PHOSPHATE-INDUCED1 (PHI1) was enhanced in the SIF2 overexpression lines but repressed in the sif2 mutant, implying that SIF2 also may regulate the calcium signaling pathway. Our results further suggest that SIF1 and SIF2 negatively regulate plant salt resistance. Taken together, we have identified a family of protein kinases that contribute to different plant stress resistance pathways. Our results shed light on the role of a specific SIF family member, SIF2, in the plant response to pathogen infection as well as its possible involvement in plant salt tolerance. Our results establish that an LRR-RLK can positively regulate plant biotic resistance but negatively regulate plant abiotic tolerance.

RESULTS

Identification of the Arabidopsis SIF Gene Family

Plant roots are the first affected tissue during osmotic stress, and they play a significant role in its perception and response. We sought to identify genes involved in the plant stress response, particularly genes that likely participate in stress signaling. We first set out to identify genes specifically or predominantly expressed in Arabidopsis root tissues using 2,904 publicly available Arabidopsis gene expression profiles conducted on an ATH1 microarray (Craigon et al., 2004). This resulted in 324 potential root-specific target genes prioritized by priority score (Supplemental Fig. S1; Wang et al., 2010), of which we specifically focused on LRR-RLKs. These filtering steps eventually led to the identification of a protein kinase, SIF1. SIF1 encodes a classic LRR-RLK protein (Fig. 1A) and is expressed predominantly in the root tissue (Fig. 1B). It is regulated intensively by drought and salt osmotic stresses (Fig. 1C), suggesting its likely involvement in the plant stress response. A phylogenetic analysis shows that SIF1 and three other Arabidopsis LRR-RLKs (SIF2–SIF4) cluster into the subfamily SIF (Supplemental Fig. S2), and their coding sequences are all closely localized on Arabidopsis chromosome I.

Figure 1.

Structure and expression analyses of SIF1. A, Structure of SIF1. PK, Protein kinase domain; SP, signal peptide; TM, transmembrane domain. B, RT-PCR analysis of SIF1 expression in 2-week-old Arabidopsis. ACTIN2 was used as the reference gene. The experiment was repeated three times. One representative image is shown. C, Two-week-old Arabidopsis seedlings grown in a hydroponic system were treated with 200 mm NaCl or drought. Root tissues were collected at the indicated time points and used in RT-PCR analysis. ACTIN2 was used as the reference gene. The experiment was repeated two times. One representative image for each treatment is shown.

SIFs are classic LRR-RLK proteins (Jones and Jones, 1997; Enkhbayar et al., 2004; Torii, 2004), as each SIF protein contains an N-terminal signal peptide with a length of 21 (SIF2–SIF4) or 28 (SIF1) amino acid residues, an extracellular LRR domain with two LRRs, a transmembrane domain, and a Ser/Thr protein kinase domain (Supplemental Fig. S3). The SIF proteins also have a high sequence similarity with each other, ranging from 73% to 86%.

SIFs Respond to Abiotic and Biotic Stresses

Our preliminary data indicate that SIF1 responds to abiotic stresses (Fig. 1). Given that SIF1 is one of four members of the SIF gene family and that the sequences of all four members are highly conserved, we speculate that the four genes would function and respond to the stresses similarly. Indeed, RT-qPCR expression analysis of the SIFs under abiotic stresses showed that the four genes all responded to salt stress (200 mm NaCl treatment) but exhibited different expression patterns. In the leaf tissue, the expression of SIF2 did not show significant change within the first 2 h but was up-regulated 4 h after salt treatment, while the expression level of SIF3 increased in the first 30 min and then declined (Fig. 2A). The expression level of SIF4 had no significant change in the leaf tissue. The transcripts of SIF1 were not detected in leaf, suggesting its root-specific expression. In the root tissue, SIF3 and SIF4 were all up-regulated dramatically, while the expression of SIF1 and SIF2 decreased progressively.

Figure 2.

Expression analysis of the SIF genes under osmotic stresses. Two-week-old Arabidopsis seedlings grown in a hydroponic system were treated with 200 mm NaCl solution (A) or dried on Whatman filter paper (B). The leaf or root samples were collected at the indicated time points and used in real-time PCR analysis. ACTIN2 was used as the reference gene. Data shown are averages of three technical replicates for two independent leaf samples or two independent root samples collected from two Arabidopsis plants grown in a hydroponic system. Error bars represent sd (n = 6). Statistically significant differences between time points were determined by one-way ANOVA. Posthoc comparisons using Tukey’s honestly significant difference (HSD) test were conducted to determine the overall difference between groups. Means not sharing the same letter are statistically significantly different (P < 0.05).

During drought stress treatment, the four genes again responded differently (Fig. 2B). In the leaf tissue, drought stress induced the accumulation of the SIF2 transcripts, but not significantly. The expression of SIF3 was repressed significantly under drought treatment. Notably, the transcripts of SIF3 were undetectable 4 h after drought treatment. SIF4 expression was significantly up-regulated 4 h after drought treatment. In the root tissue, SIF1 was down-regulated upon drought treatment, but the expression of SIF2, SIF3, and SIF4 did not show significant changes under drought treatment.

Previous studies indicated that the transcripts of SIF2 and SIF4 accumulate in leaf tissue after Arabidopsis is infected with biotrophic pathogens, such as Hyaloperonospora arabidopsidis and Pseudomonas syringae pv tomato strain DC3000 (Pst DC3000; Hok et al., 2011; Kemmerling et al., 2011). To reveal whether SIFs are involved in the pathogen resistance pathway, we first investigated the expression levels of the three leaf-expressing SIFs, SIF2, SIF3, and SIF4, in leaves infiltrated with Pst DC3000. We also used a mutant strain of Pst DC3000 (Pst DC3000 hrcC⁻), which is deficient in its type III secretion system, and two PAMP elicitors, flg22 and elf18, in the leaf treatment to test SIF responses. As shown in Figure 3, all three SIFs exhibited similar expression patterns, with the highest expression occurring 1 h after mock treatment (Fig. 3). However, when treated with pathogens or PAMPs, SIF2 exhibited different expression patterns from the mock treatment. The transcript level of SIF2 reached its peak 2 h after infiltration/inoculation of leaves with pathogens or PAMPs (Fig. 3). Notably, the expression level of SIF2 increased dramatically upon Pst DC3000 hrcC⁻ or elf18 treatment (Fig. 3), a significant induction compared with the mock treatment and the virulent pathogen Pst DC3000 infection (Fig. 3), suggesting that the effectors of Pst DC3000 may have repressed the PAMP-induced induction of SIF2. Under PAMP treatment, SIF4 exhibited a similar expression pattern to SIF2. Its expression is lower than that of SIF2, but still induced significantly (P < 0.05). Unlike SIF2, the expression of SIF3 reached its peak 1 h after mock, pathogen, and PAMP treatment (Fig. 3). These results indicate that SIF2 strongly responds to pathogens and PAMPs, suggesting its potentially important role in plant pathogen defense. Taken together, these results imply that the SIF gene family may have multiple functions and participate in both abiotic and biotic stress resistance pathways.

Figure 3.

Expression analysis of the SIF genes under pathogen and elicitor treatments. Leaves of 2-week-old wild-type Arabidopsis plants were infiltrated with 10 mm MgCl₂ as a mock control, Pst DC3000 (1 × 10⁶ colony-forming units [cfu] mL⁻¹), Pst DC3000 hrcC⁻ (1 × 10⁶ cfu mL⁻¹), 1 μm flg22, or 1 μm elf18. Leaf samples were collected at the indicated time points and used in RT-qPCR analysis. ACTIN2 was used as the reference gene. Data shown are averages of three technical replicates for two independent leaf samples collected from two Arabidopsis plants grown in different pots. Error bars represent sd (n = 6). Statistically significant differences between 0 h and the other time points were determined by Student’s t test: *, P < 0.05 and **, P < 0.01.

SIFs Exhibit Spatial and Temporal Specificity

We next investigated the expression levels of the four SIFs in different tissues of the 2- or 4-week-old Arabidopsis plants. RT-qPCR showed that SIF1 was only expressed in the root tissue of the 2-week-old plants, SIF3 and SIF4 were expressed predominantly in the leaf tissue, and SIF2 was expressed in both root and leaf tissues (Supplemental Fig. S4A). The expression patterns of the four genes changed over time with plant growth (Supplemental Fig. S4B). The expression level of SIF1 in the 4-week-old plants was quite low and could be detected only in root and rosette leaf. SIF2 was broadly expressed in nearly all the tissues of the 4-week-old plants examined, and its transcription level was the highest in root. SIF3 also showed a similar expression pattern, but with a much higher expression level than SIF2 in nearly all the tissues examined, except in the roots. SIF4 was expressed only in green tissues and flowers and exhibited a lower expression than SIF2.

GUS staining of the 2-week-old transgenic Arabidopsis plants harboring SIF1_pro/GUS, SIF2_pro/GUS, or SIF3_pro/GUS showed predominant promoter activity exclusively in the root tissue, both the root and leaf tissues, or the leaf tissue only, respectively (Fig. 4A), consistent with the RT-qPCR results (Supplemental Fig. S4A). GUS staining in the SIF4_pro/GUS plants was observed in both the root and leaf tissues (Fig. 4A), whereas RT-qPCR results showed that SIF4 was only expressed in the leaf tissue of the 2-week-old plants (Supplemental Fig. S4A), indicating that the SIF4 gene may be differentially regulated in Arabidopsis tissues at the posttranscriptional level.

Figure 4.

Activity analysis of the SIF gene promoters. Histochemical GUS staining is shown for 2-week-old (A) and 4-week-old (B) transgenic Arabidopsis plants harboring different GUS reporter systems. For each GUS reporter construct, at least two plants from two independent transformation events were stained. The photographs were taken using an optical microscope. One representative image for each sample is shown. WT, Wild type. Bars = 5 mm for (A) and Bars = 1 mm for (B).

In 4-week-old Arabidopsis plants (Fig. 4B), the histochemical staining results suggest that SIF2_pro was active in roots and leaves but weak in sepals. It did not exhibit any activity in siliques, seeds, and other parts of the flower. The spectrum of SIF4_pro activity resembles that of SIF2_pro. Obvious GUS staining in SIF3_pro/GUS plants could be observed in leaves and sepals but only slightly in root, silique, and style, correlating with the RT-qPCR data (Supplemental Fig. S4B). Unexpectedly, the activity of SIF1_pro in 4-week-old plants was very strong in the leaf and root tissues but weak in the sepals and siliques, showing a different trend from the RT-qPCR data and suggesting that SIF1 also may be regulated strictly posttranscriptionally in 4-week-old Arabidopsis plants.

SIFs Are Plasma Membrane-Anchored Proteins

As receptors in signaling transduction, LRR-RLKs are usually localized to the plasma membrane (Torii, 2004). The four SIF proteins belong to the LRR-RLK protein family, and all of them contain a transmembrane domain and a signal peptide (Supplemental Fig. S3). Therefore, we used a GFP reporter system to test the assumption that the four SIFs are all localized to the plasma membrane (Chiu et al., 1996). To this end, we produced four SIF-GFP fusion protein gene constructs. For SIF2, SIF3, and SIF4, we fused GFP downstream of their C termini. Because the full-length SIF1 cannot be expressed transiently in tobacco (Nicotiana benthamiana) leaf, we fused GFP to the C-terminal end of the first 200 amino acid residues of its N terminus, which contains the signal peptides (Fig. 5A). An additional fusion protein, PIP2A-mCherry, was composed of the membrane-anchored PIP2A protein and used as a plasma membrane marker (Fig. 5B).

Figure 5.

Subcellular localization analysis of the SIF proteins. A, Schematic diagrams of the constructs used for subcellular localization of the four SIF proteins. The DNA sequence encoding the first 207 amino acid residues in the N-terminal end of SIF1 or the full-length gene sequence encoding the SIF2, SIF3, or SIF4 protein was fused with the coding sequence of the GFP(S65T) protein and placed under the control of the cauliflower mosaic virus (CaMV) 35S promoter. B, GFP and the plasma membrane marker PIP2A-mCherry were transiently expressed in tobacco leaves as positive controls. C, SIF-GFP(S65T) was transiently coexpressed with PIP2A-mCherry in tobacco leaves. The leaf samples were examined using a Leica SPE confocal microscope. The fluorescence of SIF-GFP(S65T) is depicted in green, and the fluorescence of PIP2A-mCherry is depicted in red. Bars = 50 μm.

The SIF-GFP fusion protein and the PIP2A-mCherry fusion protein were coexpressed in tobacco leaves, and we observed green and red fluorescence emerging on the same region and overlaying perfectly with each other to emit yellow fluorescence, indicating that the four fusion proteins localized to the plasma membrane (Fig. 5C).

SIFs Play Crucial Roles in Abiotic and Biotic Stress Resistance Pathways

To understand the biological role of SIFs, we generated overexpression (OE) lines for each individual SIF family member. Semiquantitative RT-PCR analyses indicated that the four SIFs were all overexpressed successfully in the transgenic Arabidopsis plants (Supplemental Fig. S5). The SIF1 OE lines 1 and 2, the SIF2 OE lines 2 and 11, the SIF3 OE lines 1 and 5, and the SIF4 OE lines 1 and 7 were used for further analysis.

The T-DNA insertion mutants of SIFs were obtained from the Arabidopsis Biological Resource Center (Alonso et al., 2003). The positions of these T-DNA insertions in the mutant lines were confirmed by PCR (Supplemental Fig. S6A). According to the PCR results, all four SIF T-DNA insertion lines are homozygous (Supplemental Fig. S6B). The RT-PCR results indicate that the expression levels of SIFs are all reduced significantly in their T-DNA insertion lines (Supplemental Fig. S6C).

The facts that SIFs are highly conserved in their sequences and that some exhibit similar responses to abiotic or biotic stresses (Figs. 2 and 3) suggest that SIFs may be functionally redundant. To better understand their function, we decided to simultaneously repress multiple SIF genes in a single Arabidopsis mutant line. RNA interference (RNAi) technology was adopted to produce transgenic plants expressing an RNAi construct that targets a highly conserved SIF gene sequence (Supplemental Fig. S7A). RT-PCR analysis of SIF gene expression in the RNAi lines showed that only SIF1 and SIF2 were partially down-regulated (Supplemental Fig. S7B).

We first investigated the plant response to Pst DC3000 in the wild type, RNAi lines, SIF OE lines (SIF1 OE–SIF4 OE), and SIF T-DNA insertion lines (sif1–sif4). Leaves of 4-week-old plants were infiltrated with Pst DC3000 (1 × 10⁵ cfu mL⁻¹). Comparison of the plant response to pathogen infection at 3 d post inoculation revealed that the SIF2 OE line exhibited enhanced disease resistance with only a slight leaf necrosis or chlorosis, a much less severe symptom than that observed in the wild-type plants and other Arabidopsis lines. On the contrary, sif2 was more susceptible to the pathogen and displayed more severe pathogen infection symptoms than the wild type and SIF2 OE lines (Supplemental Fig. S8A). For other SIFs, no significant difference in pathogen response was observed between their OE and T-DNA insertion lines (Supplemental Fig. S8A). Similar results were obtained in 4-week-old plants spray inoculated with Pst DC3000 (5 × 10⁶ cfu mL⁻¹; Supplemental Fig. S8B), further confirming that SIF2 overexpression facilitates plant resistance to the pathogen Pst DC3000, whereas its repression compromises pathogen resistance. Collectively, these data and the previous RT-qPCR analysis showing SIF2 up-regulation upon pathogen and elicitor treatment (Fig. 3) suggest that SIF2 may be involved in the plant biotic resistance pathway and play a positive, crucial role in Arabidopsis resistance to pathogen infection.

Besides biotic stress, we also compared the plant response to salt stress in different transgenic Arabidopsis lines. We focused on SIF1, the root-specific gene, and SIF2, which is expressed in both leaf and root tissues (Fig. 2A). Two-week-old plants of the wild type, RNAi line, SIF1 OE line, SIF2 OE line, and the two T-DNA insertion mutants sif1 and sif2 were treated with NaCl and then recovered with watering. As shown in Supplemental Figure S9, the RNAi line exhibited the best performance under the salt treatment, whereas the growth of the wild-type plants and the two SIF OE lines, SIF1 OE and SIF2 OE, was much worse (Supplemental Fig. S9, A and B). These results indicate that SIF1 and SIF2 also may be involved in the plant salt resistance pathway as negative regulators. Although the phenotype analysis of the different transgenic lines under drought stress was inconclusive, preliminary data suggest differential plant drought responses in different transgenic Arabidopsis lines (data not shown).

Overexpression of SIF2 Enhances Pathogen Resistance

Our results so far strongly suggest that SIF2 may have a crucial function in pathogen defense. This prompted us to focus specifically on SIF2 to investigate its role in mediating plant pathogen defense. To this end, we furthered our analysis testing the plant response to pathogen infection by including the avirulent pathogen, Pst DC3000 hrcC⁻, in addition to the virulent pathogen Pst DC3000.

We dip inoculated 5-week-old Arabidopsis plants with 5 × 10⁸ cfu mL⁻¹ Pst DC3000 or Pst DC3000 hrcC⁻ and evaluated the pathogen development in plant leaves at 3 and 5 d post inoculation. The SIF2 OE line developed a much slighter chlorosis than the wild type and sif2 (Fig. 6A). A similar phenotype was observed in the Pst DC3000 hrcC⁻ inoculated plants (Fig. 6B). The results of bacterial titer analysis correlated with the phenotype observed, as less pathogen developed in the leaves of the two SIF2 OE plants than the wild type and sif2 (Fig. 6, C and D), indicating that overexpression of SIF2 leads to repressed pathogen development. Furthermore, we observed that the growth of Pst DC3000 and Pst DC3000 hrcC⁻ increased significantly in the sif2 leaves compared with the wild type and SIF2 OE lines, indicating an increased susceptibility of the sif2 mutant to pathogen infection (Fig. 6, C and D). Pst DC3000 hrcC⁻ is deficient in the type III secretion system, which means that only PTI will be triggered in the Pst DC3000 hrcC⁻-infected plants. The growth of Pst DC3000 hrcC⁻ was repressed in the SIF2 OE lines but increased in the sif2 plants, suggesting that the PTI response was enhanced in the SIF2 OE lines but repressed in the sif2 mutants.

Figure 6.

Phenotypic analysis of the wild type (WT), the sif2 T-DNA insertion mutant, and SIF2-overexpressing lines subjected to pathogen infection through dip inoculation. A and B, Five-week-old plants grown in soil under short-day conditions (8-h/16-h day/night) were dip inoculated with the pathogen Pst DC3000 (A) or Pst DC3000 hrcC⁻ (B). The plants were photographed 3 and 5 d post inoculation (DPI). C and D, Bacterial number in dip-inoculated Arabidopsis leaves. Leaves exhibiting symptoms were collected from Arabidopsis plants 3 and 5 d after pathogen inoculation and used for the determination of bacterial titer. Data shown are averages of four independent leaf samples collected from four Arabidopsis plants grown in different pots. Error bars represent sd (n = 4). Statistically significant differences between lines on the same day after inoculation were determined by one-way ANOVA. Posthoc comparisons using Tukey’s HSD test were conducted to determine the overall difference between groups. Means not sharing the same letter are statistically significantly different (P < 0.05).

Spray inoculation of Arabidopsis plants with 2.5 × 10⁸ cfu mL⁻¹ Pst DC3000 or Pst DC3000 hrcC⁻ gave rise to similar results (Supplemental Fig. S10). While severe symptoms developed on the leaves of the sif2 mutant plants 3 d post inoculation compared with the wild-type controls, much less chlorosis and necrosis were formed on the leaves of the SIF2 OE lines than on both wild-type and sif2 mutant plants (Supplemental Fig. S10A). Bacterial titer results also indicated that, compared with the wild-type controls, pathogen growth was enhanced in the sif2 mutants but repressed significantly in the SIF2 OE lines (Supplemental Fig. S10B).

SIF2 Regulates PAMP-Triggered Basal Immunities

Upon pathogen infection, several basal responses will be activated in the PTI pathway, including callose deposition, stomatal closure, reactive oxygen species (ROS) accumulation, expression of defense-related genes, and MAPK activation (Zipfel, 2008; Pitzschke et al., 2009; Luna et al., 2011; Daudi et al., 2012). Since SIF2 is a receptor-like protein kinase localized to the plasma membrane, it may act as a pattern-recognition receptor (PRR) in the PTI pathway, which recognizes PAMPs and triggers the downstream basal responses. To test this assumption, we investigated whether overexpression of SIF2 enhances plant basal responses triggered by the PTI pathway.

The first basal response we tested was callose deposition. Upon PTI activation, callose will be synthesized and form a matrix in the apoplast, facilitating the deposition of antimicrobial compounds that can repress pathogen growth (Luna et al., 2011). As shown in Figure 7A, no callose deposition was observed in any lines 6 h after mock treatment (Fig. 7A, top row). Upon Pst DC3000 (1 × 10⁸ cfu mL⁻¹) treatment, callose deposition was observed in all plants. However, the deposition was more abundant in the SIF2 OE lines but less so in the sif2 mutants compared with the wild-type controls (Fig. 7A, middle row). A similar phenomenon also was observed upon Pst DC3000 hrcC⁻ (1 × 10⁸ cfu mL⁻¹) treatment (Fig. 7A, bottom row). These results indicate that SIF2 regulates callose deposition.

Figure 7.

Analyses of basal immunities in the wild type (WT), the sif2 T-DNA insertion mutant, and SIF2-overexpressing lines. A, Callose deposition in Arabidopsis leaves under pathogen treatment. Leaves of 5-week-old Arabidopsis were infiltrated with MgCl₂, Pst DC3000, or Pst DC3000 hrcC⁻ for 6 h with the indicated concentrations, Aniline Blue stained, and observed under UV light. Data shown are averages of nine independent biological replicates, and two leaves were analyzed for each biological replicate. In this experiment, independent biological replicate was defined as different Arabidopsis plants grown in different pots. Error values represent sd (n = 18). Bars = 100 μm. B, Stomatal apertures of Arabidopsis leaves under Pst DC3000 treatment. Leaves of 5-week-old Arabidopsis plants were immersed in Pst DC3000 (1 × 10⁸ cfu mL⁻¹). At 1.5 and 3.5 h after treatment, stomata in randomly chosen regions on the abaxial side of the four fully expanded leaves from four plants (four leaves in total) were photographed using an optical microscope. The width of the stomatal aperture was measured using the measure function of ImageJ. Data shown are averages of four independent biological replicates each consisting of 15 stomatal apertures. In this experiment, independent biological replicate was defined as leaves collected from different Arabidopsis plants grown in different pots. Error bars represent sd (n = 60). Statistically significant differences between groups were determined by one-way ANOVA. Posthoc comparisons using Tukey’s HSD test were conducted to determine the overall difference between groups. Means not sharing the same letter are statistically significantly different (P < 0.05). C, ROS accumulation in Arabidopsis leaves under pathogen treatment. Leaves of 5-week-old Arabidopsis plants were infiltrated with MgCl₂, Pst DC3000, or Pst DC3000 hrcC⁻ for 1.5 h with the indicated concentrations. Three leaves from three plants (nine leaves in total) were then assayed for DAB staining.

Stomatal closure is another important defense mechanism triggered by PTI. Within 1 to 2 h of pathogen infection, stomata are actively closed to prevent further pathogen entry (Melotto et al., 2008). To bypass this stomatal defense mechanism, pathogens like Pst DC3000 will inject a virulence factor named coronatine to interrupt the salicylic acid/ABA-promoted stomatal closure and reopen stomata (Melotto et al., 2008; Lee et al., 2013). Since we presume that the SIF2-mediated PAMP recognition triggers basal immunities through the PTI pathway, we measured the stomatal aperture of the Arabidopsis plants under pathogen treatment to test whether SIF2 regulates the stomata-based defense. Upon Pst DC3000 treatment, a larger leaf stomatal aperture was observed in the sif2 mutant than in the wild-type controls, whereas the leaf stomatal closure was enhanced significantly in the SIF2 OE plants (Fig. 7B). A similar result was obtained when the plants were treated with Pst DC3000 hrcC⁻ (Supplemental Fig. S11). These findings suggest that SIF2 also regulates stomatal aperture, facilitating plant resistance against pathogen invasion.

ROS accumulation is an essential basal response to pathogen invasion. This basal response not only represses pathogen spread but also regulates other PAMP-triggered basal resistances, such as callose deposition and peroxidase-dependent gene expression (Daudi et al., 2012). Under the Pst DC3000 treatment, diminished 3,3′-diaminobenzidine (DAB) staining was observed in the leaves of the sif2 mutant, indicating reduced ROS accumulation in the T-DNA insertion mutant line caused by a reduced hydrogen peroxide-dependent polymerization reaction (Thordal-Christensen et al., 1997). On the contrary, a strong DAB staining was observed in the SIF2 OE plants, suggesting that ROS accumulation was strongly enhanced in the SIF2 OE plants compared with the wild-type controls (Fig. 7C). When plants were inoculated with the avirulent pathogen Pst DC3000 hrcC⁻, the sif2 mutants again exhibited a reduced ROS accumulation, whereas the SIF2 OE plants had an enhanced ROS accumulation. These results indicate that SIF2 regulates ROS accumulation (Fig. 7C).

Collectively, these results confirmed that SIF2 indeed has an essential function in regulating the basal immunities triggered by pathogen PAMPs.

SIF2 Regulates the Expression of Pathogen-Responsive Genes and Also May Participate in Calcium-Mediated Signaling during the Plant Immune Response

To further decipher the molecular mechanism of SIF2-mediated pathogen resistance, we investigated whether the altered pathogen resistance in the sif2 mutant and SIF2 OE plants is attributable to changes in SIF2-regulated defense-related gene expression. We examined the expression levels of an innate immunity marker gene, FRK1, together with a transcription factor gene, WRKY53, which is a direct target of the MAPK cassette and regulates the expression of a large number of stress-responsive genes in plants under pathogen infection and elicitor triggering (Miao et al., 2004; Zentgraf et al., 2010).

Northern-blot analysis showed that the transcripts of FRK1 could be detected 2 h after pathogen inoculation (Supplemental Fig. S12). Furthermore, a higher FRK1 transcription in the SIF2 OE plants but a lower FRK1 transcription in the sif2 mutants than in the wild-type controls were observed under the Pst DC3000 treatment (Supplemental Fig. S12). This result was further confirmed by RT-qPCR analysis (Fig. 8). Although northern-blot analysis did not seem to reveal a significant difference in FRK1 expression between the various Arabidopsis lines under the Pst DC3000 hrcC⁻ treatment (Supplemental Fig. S12), the more accurate RT-qPCR analysis revealed that the FRK1 induction 2 h after pathogen inoculation was higher in the SIF2 OE plants but lower in the sif2 mutants than in the wild-type controls upon Pst DC3000 hrcC⁻ inoculation, similar to that observed under the Pst DC3000 treatment (Fig. 8). Furthermore, FRK1 expression also was induced 2 h after inoculation with the two pathogen elicitors, elf18 and flg22, and the induction was significantly higher in the SIF2 OE plants than in the sif2 mutants (Fig. 8).

Figure 8.

Expression analysis of the defense-related genes in the wild type (wt), the sif2 T-DNA insertion mutant, and SIF2-overexpressing lines. Leaves of 5-week-old Arabidopsis plants were infiltrated with 10 mm MgCl₂ as a mock control, Pst DC3000 (1 × 10⁸ cfu mL⁻¹), Pst DC3000 hrcC⁻ (1 × 10⁸ cfu mL⁻¹), elf18 (1 µm), or flg22 (1 µm). Samples harvested at the indicated time points were used in RT-qPCR analysis to detect the transcript levels of defense-related genes. ACTIN2 was used as the reference gene. The 0-h mock control wild-type sample was set as 1. Data shown are averages of three independent leaf samples collected from different Arabidopsis plants grown in different pots. Error bars represent sd (n = 3). Statistically significant differences between the 0-h control wild-type sample and any of the other time points in various tested Arabidopsis lines under the same treatment were determined by one-way ANOVA with posthoc Tukey’s HSD test. Means not sharing the same letter are statistically significantly different (P < 0.05 or P < 0.01).

For WRKY53, northern-blot analysis results showed that, in all lines, the WRKY53 transcripts were undetectable under normal conditions but started to accumulate 30 min after the pathogen inoculation (Supplemental Fig. S12). In the early time points of infection (0–30 min), the expression of WRKY53 was only slightly different among the various lines. Unexpectedly, the transcript level of WRKY53 was higher in the sif2 mutant than in the wild-type plants under the Pst DC3000 treatment. At 1 h after inoculation of Pst DC3000 or Pst DC3000 hrcC⁻, the transcript level of WRKY53 declined rapidly in the wild-type and sif2 plants but remained elevated in the SIF2 OE lines (Supplemental Fig. S12). This corresponded with the data obtained by RT-qPCR analysis. WRKY53 induction was significantly higher in the SIF2 OE line than in the wild-type and sif2 plants upon inoculation of Pst DC3000 or Pst DC3000 hrcC⁻, and the SIF2 OE line maintained more WRKY53 transcripts than the wild-type and sif2 plants at 2 h after pathogen inoculation (Fig. 8). WRKY53 transcripts accumulated more in the SIF2 OE line than in the wild-type and sif2 plants induced by the two pathogen elicitors, elf18 and flg22, although the difference was nonsignificant (Fig. 8). Nevertheless, the RT-qPCR result showed that the transcript level of WRKY53 was higher in the sif2 mutant than in the wild-type control plants under Pst DC3000 or flg22 treatment, consistent with the observation in the northern-blot analysis.

As a secondary messenger mediating plant immune and stress response, Ca²⁺ has long been recognized to play an important role in the activation of PTI (Boudsocq and Sheen, 2013). The Ca²⁺ burst and influx induce the opening of membrane channels (e.g. influx of H⁺ and efflux of K⁺), leading to extracellular alkalization and plasma membrane depolarization (Bigeard et al., 2015). Previous studies suggest that PAMP-PRR and effector-nucleotide-binding/Leu-rich repeat immune sensor recognition can trigger the calcium influx, which regulates NPR1, RBOHD, and several Ca²⁺-dependent protein kinases (CDPKs), leading to the expression of many other downstream defense genes and a burst of ROS (Gao et al., 2014; Luo et al., 2017). To investigate whether SIF2 also participates in calcium-mediated signaling, we measured the expression of PHI1, a PAMP-induced CDPK-specific gene in the wild-type, SIF2 OE, and sif2 mutant plants subjected to pathogen and elicitor treatments (Boudsocq et al., 2010). The RT-qPCR results showed that, compared with the wild-type controls, the induction of PHI1 was enhanced significantly in the SIF2 OE line but was severely impaired in the sif2 mutant under the Pst DC3000 treatment (Fig. 8). Under the elf18 or flg22 treatment, the induction of PHI1 was repressed significantly in the sif2 mutant compared with the wild-type controls, whereas there was a similar accumulation of the PHI1 transcripts in the wild-type controls and SIF2 OE lines (Fig. 8). No significant difference in PHI1 induction was observed between the wild-type plants, sif2 mutants, and SIF2 OE lines under the Pst DC3000 hrcC⁻ treatment (Fig. 8).

SIF2 Likely Mediates Signaling Transduction through a MAPK Cassette

In the PTI pathway, the MAPK kinase modules mediate signaling transduction from the perception of PAMPs to the expression of the defense-related genes (Pitzschke et al., 2009). To investigate whether SIF2 regulates the basal immunities through the MAPK module, we examined the phosphorylation level of MPK3 and MPK6, which positively regulate pathogen resistance. For all the treatments, no phosphorylated MPK3/6 could be detected in the Arabidopsis leaves at 0 min after infiltration. A previous study revealed that MPK3/6 is fully activated 15 min after the ligand treatment (Schwessinger et al., 2011); therefore, we collected leaf samples at 5 and 15 min after infiltration to examine the MPK3/6 phosphorylation status. The results obtained showed that, after plants were treated with pathogens or elicitors, the phosphorylation level of MPK3/6 was higher in the SIF2 OE lines than in the wild-type and sif2 mutant plants (Fig. 9), indicating that SIF2 overexpression enhances the pathogen- or elicitor-triggered signaling transduction. Notably, the MPK3/6 activity in the sif2 mutant was comparable to that in the wild-type controls upon Pst DC3000 treatment but stronger than that in the wild-type controls upon flg22 treatment. This pattern correlates with the expression of the WRKY53 defense gene under the flg22 treatment (Fig. 8). On the other hand, compared with the wild-type controls, the MPK3/6 phosphorylation levels in the sif2 mutant were suppressed upon Pst DC3000 hrcC⁻ or elf18 treatment. This also largely correlates with the expression of the defense genes under the same conditions (Fig. 8). These results imply a complex SIF2-regulated MAPK signal transduction process. Together, our data demonstrate that the PR genes are indeed regulated by SIF2, and the regulation of the PR genes is probably dependent on the MAPK cassette.

Figure 9.

Phosphorylation analysis of MAPK3/6 in the wild type (WT), the sif2 T-DNA insertion mutant, and SIF2-overexpressing lines. Leaves of 5-week-old Arabidopsis plants were infiltrated using 10 mm MgCl₂ as a mock control, Pst DC3000 (1 × 10⁸ cfu mL⁻¹), Pst DC3000 hrcC⁻ (1 × 10⁸ cfu mL⁻¹), elf18 (1 µm), or flg22 (1 µm). One hundred-milligram leaf samples were harvested at 0, 5, and 15 min after infiltration and used for protein western-blot analysis to detect the phosphorylation levels of MAPK3 and MAPK6. Total protein on the polyvinylidene difluoride membrane was stained with Ponceau S dye to show approximately equal loading. The experiment was repeated twice, and one representative image for each treatment is shown. The levels of MAPKs were quantified using Ponceau S as the reference and are shown below each lane. The 5-min wild-type sample was arbitrarily set as 1. The positions of standard molecular masses of the protein and the Rubisco protein are labeled. Asterisks indicate nonspecific bands.

SIF2 Interacts with BAK1 in the Plant Response to Pathogen Infection

The plasma membrane-anchored LRR-RLK BAK1 has multiple functions in Arabidopsis. Besides the regulation of plant growth and development, BAK1 also participates in the signal transduction during pathogen invasion as a coreceptor by forming a heterodimer with other plasma membrane-localized LRR-RLKs (Postel et al., 2010; Roux et al., 2011; Schwessinger et al., 2011). We are curious about whether SIF2 interacts with BAK1 to initiate the subsequent signal transduction after it recognizes the extracellular elicitors during pathogen infection.

A bimolecular fluorescence complementation (BiFC) assay was performed to test the interaction between SIF2 and BAK1 under the pathogen treatment. SIF2, BAK1, or CHITIN ELICITOR RECEPTOR KINASE1 (CERK1) was fused separately to the C-terminal (VYCE) or N-terminal (VYNE) end of the Venus protein (Fig. 10A; Gehl et al., 2009). The empty-VYCE and empty-VYNE proteins were used as negative controls and the CERK1-VYCE and CERK1-VYNE proteins were used as positive controls to assess the efficiency of this BiFC system. The results showed that, regardless of Pst DC3000 treatment, no signal could be detected on the plasma membrane of tobacco leaves coexpressing the empty-VYCE and empty-VYNE proteins, while a strong YFP fluorescence was always detected when the CERK1-VYCE and CERK1-VYNE proteins were coexpressed (Fig. 10, B and C). On the other hand, YFP fluorescence was only detected on the plasma membrane of tobacco leaves coexpressing SIF2-VYCE and BAK1-VYNE or BAK1-VYCE and SIF2-VYNE after the infiltration of Pst DC3000 (Fig. 10D), suggesting that SIF2 and BAK1 interact with each other only under pathogen infection.

Figure 10.

Interaction of BAK1 and SIF2 under pathogen treatment. A, Schematic diagrams of the constructs used for BiFC. B, Empty-VYCE and empty-VYNE were transiently coexpressed in tobacco leaves as negative controls. C, CERK1-VYCE and CERK1-VYNE were transiently coexpressed in tobacco leaves as positive controls. D, SIF2-VYCE and BAK1-VYNE or BAK1-VYCE and SIF2-VYNE were transiently coexpressed in tobacco leaves. The leaf samples were infiltrated with or without Pst DC3000. Thirty minutes after infiltration, the leaves were examined using a Leica SPE confocal microscope. The fluorescence of Venus is depicted in red, and the chlorophyll autofluorescence is depicted in blue. Bars = 20 μm.

DISCUSSION

SIF2-Mediated Plant Response to Pathogen Infection

As classical receptor-like protein kinases, the Arabidopsis LRR-RLKs share several signature domains, including an N-terminal signal peptide, one to 32 LRR domains, a membrane-spanning region, and a protein kinase domain (Torii, 2004). Specifically, the LRR domain can identify and interact with the extracellular signaling ligand and transduce signals into cells to initiate a cellular response, conferring on LRR-RLKs the ability to perceive pathogen invasion by detecting pathogen-specific molecular patterns. In previous studies, only a few LRR-RLKs were proven to be involved in the PTI response in Arabidopsis, including FLS2, EFR, BAK1, and PEPRs (Zipfel et al., 2006; Yamaguchi et al., 2006, 2010; Chinchilla et al., 2007; Roux et al., 2011; Yamaguchi and Huffaker, 2011; Sun et al., 2013). Recently, another LRR-RLK, IMPAIRED OOMYCETE SUSCEPTIBILITY1, was identified to mediate the β-amino butyric acid-triggered PTI response (Singh et al., 2012). In our work, SIF2, another LRR-RLK, also was demonstrated to play an important role to prime the PTI response upon bacterial infection. Our data show that constitutive expression of SIF2 leads to improved plant performance against pathogen invasion, whereas the T-DNA insertion mutant sif2 is more susceptible to pathogens (Fig. 6; Supplemental Fig. S10). The enhanced resistance in the SIF2-overexpressing line is due to the enhanced basal immunities, including callose deposition, stomata closure, and ROS accumulation (Fig. 7). These enhanced basal immunities block the entry of the pathogen through the stomata and repress the development of the pathogen in the leaf tissue.

BAK1 is a multifunction LRR-RLK in Arabidopsis. Besides its critical role in the perception of brassinosteroid, previous studies showed that BAK1 also mediates PAMP perception in PTI by forming a heterodimer with FLS2, EFR, or PEPR1/2 (Li et al., 2002; Chinchilla et al., 2007; Postel et al., 2010; Schulze et al., 2010; Schwessinger et al., 2011; Sun et al., 2013), implying that BAK1 is an indispensable element in signaling transduction. Based on our BiFC results, we find that SIF2 also needs to interact and form a complex with BAK1 (Fig. 10D). This interaction between SIF2 and BAK1 depends on pathogen infection, indicating that this interaction follows the BAK1-flagellin-FLS2 model that requires a PAMP to act as the glue and stabilize the BAK1-FLS2 complex.

The MAPK modules are located downstream of PRRs. After the plasma membrane-anchored PRRs recognize PAMPs, the signal will be transmitted into the cell through the MAPK signal cascade. MEKK1-MKK4/5-MPK3/6 is thought to play a positive role in regulating plant defense (Vidhyasekaran, 2014). Previous research indicated that pathogen infection and the PAMP elicitors, such as flg22 and elf18, induce strong MPK3/6 phosphorylation, which positively regulates the downstream basal responses (Takahashi et al., 2007; Beckers et al., 2009; Pitzschke et al., 2009; Meng et al., 2013). In this study, we observed that the phosphorylation level of MPK3/6 in the SIF2 OE plants was higher than that in the wild-type and sif2 plants upon pathogen or elicitor treatment, suggesting that SIF2 also may utilize the MPK3/MPK6 signaling pathway to regulate plant defense (Fig. 9). Overexpression of SIF2 may enhance the MPK3/MPK6-mediated signaling transduction, especially under the treatments of the two pathogen strains, Pst DC3000 and Pst DC3000 hrcC⁻, and the elicitor, flg22, resulting in more intensive basal immunities in the SIF2 OE plants than in the wild-type and sif2 mutant plants. We also noticed that the MPK3/6 activity in the sif2 mutant was lower than that in the wild-type controls under Pst DC3000 hrcC⁻ or elf18 treatment, comparable to that in the wild-type controls under the Pst DC3000 treatment but stronger than that in the wild-type controls under the flg22 treatment (Fig. 9). These phenomena imply that SIF2 may negatively regulate signaling transduction in the early stage of pathogen infection, making our hypothetic SIF2-MEKK1-MKK4/5-MPK3/6-resistant gene model more complex.

The activated MPKs induce the expression of defense-related genes through activating transcription factors (Gómez-Gómez and Boller, 2002; Pitzschke et al., 2009). The WRKY transcription factors are DNA-binding proteins that can recognize and bind to the cis-regulatory elements in the promoter region of the functional genes, regulating their expression at the transcriptional level. Under pathogen infection or salicylic acid treatment, 49 out of 72 WRKY mRNA levels are altered, indicating that they are important components involved in the pathogen defense mechanisms (Dong et al., 2003). WRKY53 is identified as both a positive and a negative regulator of the basal responses and targets at least seven other WRKY proteins, including two direct MAPK targets, WRKY22 and WRKY29, suggesting that it is a centerpiece of plant defense signaling transduction (Miao et al., 2004; Zentgraf et al., 2010). In this study, northern hybridization and RT-qPCR analysis both indicated that, among the different Arabidopsis lines tested, the SIF2 OE lines had the highest WRKY53 expression after pathogen treatment (Fig. 8; Supplemental Fig. S12) and exhibited the strongest basal defenses (Fig. 7). This finding demonstrates that overexpression of SIF2 enhances the expression of WRKY53 at a later time point of pathogen infection, which then directly (induction of cell senescence) or indirectly (through other WRKY protein networks) induces strong basal defenses against the pathogen. Miao et al. (2007) showed that MEKK1 could interact directly with WRKY53 and induce its expression, implying that SIF2 also may be involved in plant defense through the SIF2-MEKK1-WRKY53-resistant gene signaling pathway (Miao et al., 2007).

Similar to MAPK western analysis, we noticed that there were more WRKY53 transcripts accumulating in the sif2 mutants than in the wild-type controls in the early stage of pathogen (Pst DC3000) infection or flg22 treatment (Fig. 8; Supplemental Fig. S12). Further study is necessary to explain this phenomenon. Gene expression analysis also suggests that the regulation of FRK1 was affected by the altered expression of SIF2 (Fig. 8; Supplemental Fig. S12). FRK1 expression was lower in the sif2 mutant but was strongly enhanced in the SIF2 OE lines. FRK1 is a receptor-like protein kinase gene. Its increased expression implies that the innate immunity system has been activated. Compared with the wild-type controls, the enhanced expression of FRK1 in the SIF2 OE lines, the reduced expression in the sif2 mutant lines, and the fact that pathogen infection triggers the SIF2-BAK1 interaction (Fig. 10D) suggest that SIF2 plays an important role in the innate immunity system for disease resistance. Furthermore, the altered expression of PHI1 in the SIF2 OE lines and sif2 mutants under the pathogen (Pst DC3000) infection and elicitor (elf18 and flg22) treatment suggests an impacted calcium-mediated signaling in the sif2 T-DNA insertion mutant and SIF2 overexpression lines, implying that SIF2 may be involved in early events of the PTI pathway, including the regulation of calcium influx.

Taking these results together, we propose a hypothetic model showing how SIF2 is involved in the pathogen-triggered signaling pathway (Fig. 11). Briefly, upon pathogen infection, SIF2 binds the PAMP elicitor, resulting in the formation of the SIF2-elicitor-BAK1 complex. MEKK1 is phosphorylated by the activated kinase domain of BAK1 and/or SIF2, leading to the activation of MKK4/5 and, finally, MPK3/6. The active MPK3/6 then interacts with the downstream WRKY protein(s), which positively regulate(s) WRKY53. The phosphorylated MEKK1 also may interact with WRKY53 directly. WRKY53 and other possibly involved WRKY proteins induce the expression of resistance genes, ultimately leading to the activation of basal immunities, including callose deposition, ROS accumulation, and stomatal closure. The SIF2-involved PTI pathway also may regulate calcium influx, inducing the expression of pathogen resistance genes through the calcium-mediated signaling pathway via CDPKs, such as PHI1. It also should be noted that, since SIF2 is an RD (Arg-Asp) RLK, not a non-RD RLK like the other well-studied PRRs, it cannot be ruled out that SIF2 may act as a coreceptor for PAMP elicitor binding in the PTI pathway. In this hypothetical model, many questions remain unknown. For example, what is the specific PAMP elicitor recognized by SIF2? Since SIF2 is an RD RLK, not a non-RD RLK like the other well-studied PRRs, how exactly does the SIF2-BAK1 complex work? How is the SIF2-mediated signal transduced through the MAPK cascade to the WRKY53 protein?

Figure 11.

Model of the SIF2-mediated signaling pathway. Upon pathogen infection, SIF2 binds a PAMP elicitor, resulting in the formation of the SIF2-elicitor-BAK1 complex. MEKK1 is phosphorylated by the activated kinase domain of BAK1 and/or SIF2, leading to the activation of MKK4/5 and, finally, MPK3/6. The active MPK3/6 then interacts with the downstream WRKY protein(s), which positively regulate(s) WRKY53. The phosphorylated MEKK1 also may interact with WRKY53 directly. WRKY53 and the other possibly involved WRKY proteins then induce the expression of resistance genes, ultimately leading to the activation of basal immunities, including callose deposition, ROS accumulation, and stomatal closure. Besides the SIF2-mediated signaling pathway, the other PRRs, such as FLS2 or EFR, also form a heterodimer with BAK1 and activate basal immunities through a similar signaling pathway. The SIF2-involved PTI pathway also may regulate calcium influx, and the increased amount of cytosolic calcium could induce the expression of pathogen resistance genes through the calcium-mediated signaling pathway via CDPKs, such as PHI1. The basal immunities also could be regulated by other mechanisms. For example, under the virulence pathogen infection, the ABA pathway could promote rapid stomatal closure. The verified steps and elements in this scheme are highlighted in red.

We currently do not have evidence showing that SIF2 is involved physically in flg22 or elf18 perception. The results of the RT-qPCR (Figs. 3 and 8) and western-blot (Fig. 9) analyses demonstrated that SIF2 may be involved indirectly in flagellin and/or ELF perception, as treatment with the elicitors (flg22 and elf18) resulted in a significant induction of SIF2. Some of the PTI responses (the expression of defense-related genes and the phosphorylation of MAPKs) were decreased in the sif2 T-DNA insertion mutant (especially the expression of PHI1 in response to both flg22 and elf18 and the phosphorylation of MAPK3/6 in response to elf18) but enhanced in the SIF2 OE line (especially the expression of FRK1 in response to elf18 and MAPK3/6 phosphorylation in response to flg22). Similarly, a previous study showed that the absence of the β-amino butyric acid receptor LECTIN RECEPTOR KINASE VI.2 could negatively impact the response of Arabidopsis to flg22 and elf26 (Singh et al., 2012). Both our data and the previous studies suggest that PRRs may work closely with each other, cooperatively coordinating the plant immune response. The absence of one PRR may impact the functions of other PRRs. Further study would allow a better understanding of whether and how SIF2 is involved in the perception of flagellin or ELF.

Our research here suggests that the phosphorylation level of MPK3/6 and the expression level of WRKY53 are both enhanced in the SIF2 OE plants. However, it was demonstrated previously that WRKY22, but not WRKY53, is the direct substrate of the activated MPK3/6 (Asai et al., 2002; Pitzschke et al., 2009). A recent study indicated that WRKY53 might be regulated by WRKY22 (Hsu et al., 2013). In addition, WRKY53 was proven to target many other WRKY proteins, including WRKY22 and WRKY29 (Miao et al., 2004). These studies imply that, in the SIF2 signaling pathway, MPK3/6 may regulate the WRKY53 protein by activating WRKY22, and then the activated WRKY53 and WRKY22 regulate each other to amplify the signal.

The Multifunctional SIF Gene Family

The SIF kinase family, like the SOMATIC EMBRYOGENESIS RECEPTOR KINASE family, which is involved in both the brassinosteroid perception signaling pathway and the signaling transduction of plant defense, may play multiple roles in different Arabidopsis resistance mechanisms (Hecht et al.,2001; Albrecht et al., 2008; Roux et al., 2011; Gou et al., 2012). Although all four SIF proteins are plasma membrane-anchored proteins and have similar structures (Fig. 5C; Supplemental Fig. S3), their expression patterns are distinct from each other (Fig. 4; Supplemental Fig. S4). Additionally, SIF1 to SIF4 are differentially regulated under abiotic stresses and biotic stresses (Figs. 2 and 3). In later experiments, we found that SIF1 and SIF2 are negative regulators of salt resistance (Supplemental Fig. S9), and SIF2 also is a critical positive regulator in the pathogen defense mechanism. All these results suggest that the SIF gene family is a multifunctional kinase family. SIFs may recognize different ligands and elicitors and, therefore, are involved in different resistance mechanisms responding to different stresses. The next step would be to verify the functions of SIF1 and SIF2 during the plant salt response and unravel the roles of SIF3 and SIF4. SIF3 and SIF4 are strongly expressed in green tissues, especially in leaves (Supplemental Fig. S4), suggesting their important functions in the aerial part. A machine learning-based analysis revealed that SIF4 was regulated intensively under salt, drought, and wound stresses (Ma et al., 2014). Both our work and a previous study showed that SIF4 is strongly up-regulated in Arabidopsis subjected to Pst DC3000 treatment (Fig. 3; Hok et al., 2011), which gives clues about SIF4 function.

CONCLUSION

To this end, our work focused on deciphering the signaling pathway in plant responses to both abiotic and biotic stresses. We identified a protein kinase family in Arabidopsis composed of four members (SIF1–SIF4) that function as receptors on the plasma membrane. The evidence from our work indicates that SIF2, one of the SIF kinase protein family members, plays a critical role in the bacterial resistance pathway. SIF2 functions as a PRR, sensing a pathogen’s presence and interacting with the coreceptor BAK1 to transmit the signal to the cytoplasm and activate downstream defense-related genes and basal immunities through the MAPK cascade. Our work also shows that SIF1 and SIF2 may negatively regulate the salt resistance of Arabidopsis.

MATERIALS AND METHODS

Growth Conditions of Plants and Bacteria

For abiotic stress experiments, Arabidopsis (Arabidopsis thaliana) plants were grown in soil under a 16-h-day/8-h-night photoperiod at 22°C day/20°C night in a growth chamber. For quantitative analysis of gene expression under abiotic stresses, Arabidopsis plants were grown in a hydroponic system under a 16-h-day/8-h-night photoperiod at 22°C day/20°C night in a growth chamber (Huttner and Bar-Zvi, 2003). For biotic stress experiments and quantitative analysis of gene expression under biotic stresses, Arabidopsis plants were grown in soil under an 8-h-day/16-h-night photoperiod at 22°C day/20°C night in a growth chamber.

For the biotic stress experiment, Pseudomonas syringae pv tomato strain DC3000 and Pst DC3000 hrcC⁻ were grown in King’s B liquid medium with rifampin for 24 h at 28°C (King et al., 1954). The bacterial culture was then centrifuged, and the pathogen cells were resuspended in 10 mm MgCl₂ to the desired densities.

The Escherichia coli strain DH5α was used in this study. The Agrobacterium tumefaciens strains used in this study are LBA4404 and GV3101.

DNA and RNA Isolation, cDNA Synthesis, and RACE

Plant genomic DNA was isolated following a previously described method (Zhou et al., 2013). Plant total RNA was isolated with Trizol reagent (Ambion) from 100-mg plant samples according to the manufacturer’s instructions. The first-strand cDNA was synthesized following a previously described method (Yuan et al., 2016). The 5′ end and 3′ end cDNA fragments of SIF1 and SIF2 were amplified with the SMARTer RACE 5′/3′ commercial kit (Clontech) following the manufacturer’s instructions.

Quantitative Analysis of Gene Expression by Northern Blotting and RT-qPCR

For northern analysis, 15 μg of total RNA denatured at 95°C was separated on a 1% agarose formaldehyde gel, transferred to Hybond-N⁺ nylon membrane (Amersham), and hybridized with the PCR-amplified and radiolabeled probes (300–400 bp).

RT-qPCR was performed with the SYBR Green Supermix (Bio-Rad) and the iQ5 real-time detection system (Bio-Rad) according to the manufacturer’s instructions. The RT-qPCR results were determined by using the ΔΔCt method (Yuan et al., 2016).

The primers used in northern blotting and RT-qPCR are all listed in Supplemental Table S1.

Protein Extraction, SDS-PAGE, and Western Analysis

Plant total protein extraction and SDS-PAGE (Merck Millipore) were conducted as described previously (Conrath et al., 1997). Protein was then transferred to a polyvinylidene difluoride membrane using the iBlot2 western-blot transfer system (Thermo Fisher Scientific). Western analysis was performed using Phospho-p44/42 MAPK (Erk1/2; Thr-202/Tyr-204) rabbit monoclonal antibody (Cell Signaling Technology) as primary antibody at a dilution of 1:800 in 5% (w/v) BSA Tris-buffered saline + Tween and goat anti-rabbit IgG (H+L) Poly-HRP (Invitrogen) as secondary antibody at a dilution of 1:3,000 in Tris-buffered saline + Tween. Signal was developed with Clarity Western ECL Substrate (Bio-Rad) and detected using Chemidoc XRS+ imager (Bio-Rad).

Plasmid Construction and Plant Transformation

For the histochemical GUS staining experiment, the predicted 2,078-bp SIF1 promoter region, the predicted 828-bp SIF2 promoter region, the predicted 1,524-bp SIF3 promoter region, and the predicted 1,141-bp SIF4 promoter region were amplified from Arabidopsis genomic DNA with iProof high-fidelity DNA polymerase and subcloned separately into the binary vector, p35S/bar/nos-GUS/nos, in the upstream part of the GUS gene, resulting in p35S/bar/nos-SIF_pro/GUS/nos.

To produce GFP-SIF fusion constructs, the cDNA encoding the first 207 amino acid residues of the SIF1 N-terminal region and the full-length cDNAs (without stop codon) of SIF2, SIF3, and SIF4 were cloned from cDNA with iProof high-fidelity DNA polymerase and subcloned separately into the pCambia binary vector, p35S/C4ppdk1/sGFP(S65T)/nos-p35S/hptII/nos, upstream of sGFP(S65T).

To produce SIF1 and SIF2 overexpression constructs, the full-length cDNAs of SIF1 and SIF2 were cloned from cDNA with iProof high-fidelity DNA polymerase and subcloned separately into the pCambia binary vector under the control of the cauliflower mosaic virus 35S promoter, resulting in p35S/SIF1/nos-p35S/hptII/nos and p35S/SIF2/nos-p35S/hptII/nos.

For RNAi plasmid construction, a 320-bp DNA fragment that was highly conserved across the whole SIF gene family was amplified from cDNA and subcloned into the binary vector, forming an RNAi construct with the expression cassette, rice (Oryza sativa) Ubi promoter/SIF (antisense)/3′ GUS linker/SIF (sense)-p35S/hptII/nos.

A. tumefaciens-mediated Arabidopsis transformation was conducted according to the previously described method (Clough and Bent, 1998).

The primers used in plasmid construction are all listed in Supplemental Table S1.

Histochemical GUS Staining

GUS activity was assayed by histochemical staining as described previously (Luo et al., 2000).

Detection of ROS Accumulation

Leaf samples were collected and vacuum infiltrated with 1 mg mL⁻¹ DAB solution (pH 3.8), followed by incubation in the dark for 14 h at room temperature. The samples were then destained in 90% (v/v) ethanol at 70°C until chlorophyll was completely removed and stored in 70% (v/v) ethanol.

Measurement of Callose Deposition, Stomatal Aperture, and Bacterial Titer

Callose deposition, stomatal aperture, and bacterial titer were determined following the previously described methods (Tsai et al., 2011; Singh et al., 2012).

Subcellular Localization and BiFC

Subcellular localization and BiFC were performed as reported previously (Sparkes et al., 2006; Gehl et al., 2009).

Statistical Analysis

One-way ANOVA [F(df_between, df_within) = F ration, P = P value, where df = degrees of freedom] with posthoc comparisons using Tukey’s HSD test was conducted to determine the statistically significant difference between the means from groups. Means not sharing the same letter are statistically significantly different (P < 0.05).

Accession Numbers

Sequence data from this article can be found in the Arabidopsis Genome Initiative and European Molecular Biology Laboratory databases under the following accession numbers: SIF1 (AT1G51840 and AT1G51830), SIF2 (AT1G51850), SIF3 (AT1G51805), SIF4 (AT1G51820), ACTIN2 (AT3G18780), FRK1 (AT2G19190), WRKY53 (AT4G23810), rRNA 18S (X16077), MPK3 (AT3G4564), MPK6 (AT2G4379), and PHI1 (AT1G35140).

The germplasm information for the Arabidopsis T-DNA insertion mutants can be found in The Arabidopsis Information Resource under the following accession numbers: sif1 (SALK_046053C), sif2 (SALK_068030), sif3 (SALK_051190), and sif4 (SALK_055952).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Genome-wide identification of root-specific genes.

• Supplemental Figure S2. Phylogenetic analysis and genomic organization of the SIF genes.

• Supplemental Figure S3. Alignment of the SIF proteins.

• Supplemental Figure S4. Expression analysis of the SIF genes in different tissues of Arabidopsis.

• Supplemental Figure S5. Overexpression of SIF genes in Arabidopsis.

• Supplemental Figure S6. Analysis of the SIF T-DNA insertion mutants.

• Supplemental Figure S7. Construction and gene expression analysis of RNAi lines.

• Supplemental Figure S8. Phenotypic analysis of different Arabidopsis lines under pathogen infection.

• Supplemental Figure S9. Phenotypic analysis of different Arabidopsis lines under salt treatment.

• Supplemental Figure S10. Phenotypic analysis of the wild type, the sif2 T-DNA insertion mutant, and SIF2-overexpressing lines subjected to pathogen infection through spray inoculation.

• Supplemental Figure S11. Stomatal apertures of the Arabidopsis leaves under Pst DC3000 hrcC⁻ treatment.

• Supplemental Figure S12. Northern analysis of the defense-related genes in the wild type, the sif2 T-DNA insertion mutant, and SIF2-overexpressing lines.

• Supplemental Table S1. Primers used in this research.